If you’ve been following the news lately, you’ve seen that major tech companies are making dramatic workforce reductions by deploying AI tools. The tech company Block recently laid off 4,000 employees, citing that intelligent tools allow it to operate more efficiently with much smaller teams.
While decisions like these are becoming aspirational for large businesses trying to become leaner, it led me to reflect on something important:
What can small businesses — who already operate with small teams — do to become more effective with AI?
That question is why I’ve written these tips to help you understand the do’s and don’ts of using AI in your small business.
Most advice about AI is built on a simple premise: automate everything you can, as fast as you can. But for small businesses, that approach carries a serious risk — eroding what makes customers choose you over a larger competitor.
Your advantage as a small business is trust, personality, and the relationships you build with your customers. Understanding where to draw the line with AI will save you money, protect your reputation, and help you use AI where it genuinely makes your life easier.
That’s why I’ve outlined three areas you should avoid using AI in — at least for now.
The Three Danger Zones for Small Business AI
1. Using AI-Generated Images
Research has found that 54% of people can correctly identify an AI-generated image when shown one, according to published perception studies. More importantly, consumers consistently rate AI-produced content as less authentic and less reliable than content created by humans.
The reputational stakes are real. A survey cited in brand perception research found that 33% of respondents said AI-generated content worsens their perception of a brand.
And it’s not just a problem for small players. Brands like McDonald’s and Coca-Cola have faced public backlash for holiday advertising campaigns that audiences described as “creepy” or emotionally hollow — a phenomenon researchers sometimes call the Uncanny Valley effect when applied to brand communications.
For a small business, the math is even less forgiving.
A major corporation can survive a wave of negative social media comments about a holiday ad. A craft vendor at a local market — or a boutique studio with 400 Instagram followers — may not recover from the same moment of misplaced automation.
2. Customer Service: Where AI Hallucinations Become Legal Liabilities
Customer service feels like an obvious place to deploy AI. It’s repetitive, time-consuming, and expensive.
But the data on how customers actually feel about AI-powered service offers a useful reality check. Research from U.S. consumer surveys consistently finds that roughly 45% of adults report a negative experience with AI chatbots, with complaints centered on an inability to handle nuanced questions and a lack of empathy in difficult situations.
The consequences can go beyond a frustrated customer.
One of the most widely cited cases in AI governance discussions is Air Canada, which was held legally liable after its AI chatbot provided a passenger with incorrect information about a bereavement fare policy. A Canadian tribunal ruled that Air Canada could not disclaim responsibility for statements made by its own automated system — even though the chatbot had effectively invented the policy.
For a small business, that kind of liability is not theoretical.
If your chatbot tells a customer the wrong refund window, quotes an incorrect price, or misrepresents what a service includes, you own that conversation — legally and reputationally.
The time you save may not be worth what it costs when something goes wrong.
3. Replacing People: What Early Adopters Are Learning the Hard Way
Perhaps the most notable AI failure story of the past two years is Klarna. The Swedish fintech company became a case study for aggressive AI adoption after announcing it had used AI to replace a significant portion of its customer service workforce.
The results were reported as a major success — until the company quietly began redeploying staff from marketing, engineering, and legal departments to answer customer calls, according to reporting by Business Insider.
Klarna is not alone.
According to a June 2025 prediction from Gartner, by 2027, half of organizations that expected to significantly reduce their customer service workforce through AI will abandon those plans, as the goal of fully integrated AI workers proves unachievable in practice.
A Gartner survey of 163 customer service leaders found that 95% plan to retain human agents specifically to define and oversee AI’s role — a “digital first, but not digital only” approach.
As Gartner Senior Director Analyst Kathy Ross put it:
“While AI offers significant potential to transform customer service, it is not a panacea. The human touch remains irreplaceable in many interactions.”
If Fortune 500 companies with dedicated AI teams are discovering this, it’s worth taking seriously as you evaluate what roles AI can realistically fill in your business.
The One Principle That Ties This Together
Underneath all three of these failure patterns is the same root cause:
People are using AI for tasks that require consistent, reliable, human-feeling outputs — and AI is not built for that. Not yet.
Here is the simplest way to think about it:
AI is a very confident guesser.
For many tasks, a confident guess is perfectly fine — even great.
But for tasks where getting it wrong has real consequences — your pricing, your customer promises, your brand identity, your legal obligations — a guess is the last thing you want.
This is why AI tends to work best in what you might call back-of-house roles: the operational, administrative, and research tasks that happen before you interact with a customer or deliver your product.
Think:
Drafting internal documents
Researching new suppliers
Summarizing long reports
Generating a first draft that a human will review and refine
These are tasks where a confident guess saves you time without putting anything at risk.
A Simple Test You Can Use Tomorrow
Before using any AI tool in your business, run it through these three questions:
1. Does this task take more time than I would like it to?
2. Is this task not core to my business, or not a unique aspect of how I operate?
3. Have I wanted to hire someone to do this task but could not find the budget?
If the answer is yes to all three, and the task does not overlap with the three danger zones outlined above — customer-facing creative work, live customer service, or roles where errors create legal or reputational exposure — then you likely have a strong candidate for AI assistance.
Start there.
Build your confidence with low-stakes wins.
Then expand from that foundation as you develop a clearer sense of where AI adds genuine value in your specific context.
A Final Thought
There is no shortage of thought leadership about AI for large enterprises. Every consulting firm, analyst house, and tech company has a framework for how Fortune 500 companies should integrate AI into their operations. That content is useful — for the audience it was written for.
But small businesses operate with different constraints, different risk tolerances, and different competitive advantages.
What works at scale often backfires in a context where every customer relationship is personal and every misstep is visible.
My goal is to help small business owners use innovative technology in ways that make running a business more sustainable — not more complicated.
If you’re interested in learning more, please reach out. I’m always happy to help.
Sources & Further Reading
Gartner (June 2025): “Gartner Predicts 50% of Organizations Will Abandon Plans to Reduce Customer Service Workforce Due to AI” — gartner.com
Gartner (September 2025): “Gartner Predicts None of the Fortune 500 Companies Will Have Fully Eliminated Human Customer Service by 2028”
Gartner (December 2025): “Gartner Survey Finds Only 20% of Customer Service Leaders Report AI-Driven Headcount Reduction”
Air Canada chatbot liability ruling, Canadian Civil Resolution Tribunal, 2024
Klarna AI reversal reporting, Business Insider, 2024–2025
Consumer perception of AI-generated content: Brand trust and authenticity research (multiple published studies, 2023–2024)
Google Cloud: “101 Gen AI Use Cases with Technical Blueprints” — cloud.google.com
CBC: “Block lays off 4,000 workers citing AI. But how much does AI actually have to do with it?” — cbc.ca